Endosymbiont: Configuration guide to your virtual interface implant

Virtual Interface Implants (VII) are brain implants that interface bio-based life with cybernetics. They are used for connectivity to networks and as computing platform. More often, the bulk of their computing cycles is allocated to an AI. Biological life can become highly dependent to the embedded AI, while the AI retains mobility. It is thus best to keep the AI from gaining too much free will.

Unboxing a VII

The VII is implanted in the cortex and has an input/output to all of the sensory systems. This means that it reads and writes to the host's senses by default. It can read surface thoughts at the subvocalization level.  "Puppet implants" have write access to skeletal muscles, but there are very few reasons why anyone would agree to grant this level of permission to the body that is host to a free thinking mind. Furthermore, most VII have biomonitors and can keep tabs on vital signs and general health. Some may even provide prophylactic and remedial treatments.

The VII is otherwise akin to a very smart phone. It has an absolute sense of direction for as long as it can read GPS signal and or WiFi. It has a wireless data and a fibre optic port usually coming from behind the ear. The fibre optic port is unjammable and more discreet and allows for higher bandwidth.

Computing cycles

Edwinn got his implant through the corporate route and has a more powerful processor system installed (Small computer, compact, Complexity/6). The standard tiny computer has a Complexity/5, which would not be sufficient to run his low-sapiens AI (or LAI, or infomorph). The base level of empathy and companionship would be lost on a non-sapiens AI running at Complexity/4. The dedicated controls on the VII allows both the AI and the host to operate the device simultaneously.

As described in B472, the Complexity/6 VII can run a Complexity/6 infomorph and another Complexity/6 software process (see below on software agents). This would be a bad idea, however, as it would preclude the running of lower level processes (such as telepresence/4 and augmented reality/4). A better strategy is to stagger the resources to: 1 Complexity/5, 9 X Complexity/4 and, 10 X Complexity/3 software. Although alternative configurations exist, this is a good rule of thumb.

Superhumanity through software configuration

This is where things get tricky. Let's just get into the thick of it: 

1. A host has its own sets of skills that can be used normally. 
2. IQ-based skills can be enhanced by software agents other than the infomorph. Agents are often required to operate IQ-based skills at TL9 and above. With the VII, no need for an external device. For the sake of simplicity, let's assume that it is possible to acquire minimal agent via the cloud for all IQ-based skills, all at Complexity/2 (single point)
3. An informorph has its own skills that it learns on its own. IQ-based skills still need and agent at TL9 or above. DX-based skills can only be used if the infomorph has puppet controls (or is teleoperating a cybershell).  Keep in mind that AIs have visualization: given a minute to analyse and simulate courses of action, the informorph may acquire up to +4 on a skill check.
4. An infomorph has a configurable Modular Ability (computer Brain) where it is possible to acquire character points in skills and some mental advantages through software. This is very flexible, but has a -3 on skills check in stressful/fast-paced situations because it is not an inherent skill.

Selecting the AI that is right for you

The nonsapien AI (NAI) - With a base IQ of 8 (Complexity/4) and the Automaton metatrait, expect the NAI to perform its task with no hope to take initiative or interest to its surrounding and broader context. Nearly nowhere the NAI is considered to be living, with any claim to rights. 

The low-sapiens AI (LAI) - With a based IQ of 9 (Complexity/6), the LAI shows a great level of initiative and holistic thinking. They are treated as slaves in most jurisdiction, and equivalent to children in the Caliphate. Their low-empathy makes them hard to faze, staid but deeply honest. LAI make better general purpose and personal assistant than NAI. The loyalty of an LAI wouldn't be a hard thing to acquire and maintain.

The sapiens AI (SAI) - With a base IQ of 9 and Complexity/7, the SAI employs human-like structures to store information and emotions. They understand humans very well, some may say too well. They are, however, program to be honest and inherently mentally stable.  Maintaining a SAI in a VII is very much akin to getting married as it is a partnership between two minds. In fact, SAI rarely can run on VII hardware nor would they want to become anyone's pet or servants. The EU and the Caliphate are actually granting full rights to SAI as voting citizens.

Configuring Modular Abilities (Computer Mind)

This is an advantage intrinsic to an AI (infomorph). It is a pool of CP that can be configured. Its point values is the following: 6+(4*Number of configurable cps). These are software-driven tools that boosts the skills of the infomorph. To fully grok how this works with skills, let's keep in mind that Complexity/3 grants 2cp, Complexity/4 grants 4, Complexity/5 grants 8cp and Complexity/6 grants 16cp. The cost is 100Cr per cp in the pool.

Here are some typical loadouts:

• Modular Abilities (Computer Brain) 3 (2,2,2) [30] - allows three Complexity/3 skills to be active at one time. Complexity/3 skills are useful to be considered trained in untrained skills. The 2cp doesn't otherwise provide any support for informorph already with 4cp in a particular skill. $600.
• Modular Abilities (Computer Brain) 3 (Eidetic, 2) [34] - This loadout grants the infomorph eidetic memory for all sensory inputs, and some skill support. The caveat for this setup is that the high-resolution memory of the host is owned by the infomorph. $900.
• Modular Abilities (Computer Brain) 4 (4) [22] - A lower cost expert boost for an infomorph. $400.

Software agents

IQ-based skills at TL9 and above must use basic software. Basic software and are assumed to have such low complexity that we don't need to keep track of their resource requirements. As per Ultra-Tech p.25, a good quality agent grants +1 to a skill are Complexity/4 for E skills ($30), and Complexity/5 for others ($100). Fine quality agents are the top of the line and grant +2 on skills! Their complexity and costs for Easy and other skills are respectively Complexity/6 ($300) and Complexity/7 ($1000). 

As you can see, even Edwin with his souped-up VII could only get away with running 1 good-quality agent for an Average or Hard difficulty skill. 

Most importantly, software agents are usable by either the host or the infomorph. A Good-quality agent used for an untrained skills, however, only slightly change the penalty of using the skill at default. The right tool in naive hands isn't a replacement for training.

In Action...

Edwin wants to be promoted from Rank 1 to Rank 2 in the company. To do this, he needs to petition a Rank 3 executive as his immediate supervisor won't be sympathetic. He needs to succeed an influence check with a -2 penalty. His Administration skill is a mere 9, improved to 10 due to an agent running on the VII. His infomorph to prepare the terrain by filing a number of reports at the right place. Maus has Admin-8, boosted to a 9 by its modular abilities, with visualization, this may climb up to further (it fails the IQ by one, no visualization bonus). 

Maus scans the office's communications and gather a series of facts that Edwin can use (Pass its admin check). Edwin, with this support from its infomorph, has now Admin-11, made Admin-9 dues to the differential in ranks. He fails by 5 while the executive succeeds by 1. Edwin's request doesn't even get a reply. Just asking for it isn't going to be good enough.